Two Years of AI in Marketing: What Actually Changed

2 September 2026

In late 2024, AI could answer questions and computer use was a beta curiosity. Two years on — through Gemini 2 and 3, Claude 3.5 to today's frontier models, GPT-5 and its successors, and Copilot spreading through every Office install — it's worth separating what actually changed from what merely trended.

What really changed

• Search broke its promise. The click-per-question economy that funded twenty years of content marketing has been rewritten by AI answers. Visibility now means being cited and remembered, not just ranked.
• Content stopped being scarce. Words, images and video became effectively free to produce — so attention repriced everything. Generic content is worthless at any volume; genuine content commands a premium.
• AI became a colleague, not a toy. The shift from chatbots to agents means the useful question is no longer "what can it write?" but "what can I delegate — and how do I check it?"

What didn't change

People buy from businesses they know, like and trust. Owned channels — your email list, your customer data, even the humble letterbox — outlasted every algorithm change, and the industry data says they're performing better than ever. And the skills gap, not the technology gap, still decides who benefits: the same tools sit on everyone's desk, doing brilliant work for some and gathering dust for most.

Our bet for the next two years is the same as our name: the winners will use artificial intelligence to become more human-facing, not less. Be intelligent. Don't be artificial.